#

OP API Documentation PREVIEW

Build your own algorithm. The new OP API allows you to create dynamic feeds from different, independent websites with complex filtering and grouping methods. Make custom JSON or RSS feeds based on phrases, hashtags, webring membership, originating or target url, with date filtering, fuzzy or exact matching, and the ability to exclude any of the same types of values.

BACKLINKS

DRAFT DO NOT SHARE

Overview

The Octothorpes Protocol API is about getting statements out of the database.

Common use cases include:

simply following a hashtag
getting a list of all links to your site
following all new posts across a webring you belong to
following all the posts from a group of unrelated urls that use the same hashtag (even if they each spell it slightly differently)
creating a live-updating blogroll or list of members of a webring

Worth mentioning is that OP is not a search engine or a crawler. OP servers only index pages that actively request it, so it is so far unaffected by many common issues with low-quality results from contemporary search engines.

How do I get data in?

You create or update statements by asking an OP server to index a webpage you own with markup on it that can be recognized by the server. See the quickstart guide or the current, slightly outdated documentation on how to do that.

Why is it out of date? OP now uses the concept of a Harmonizer to allow you to define your own methods of adding octothorpes to your website. This means you don’t necessarily have to change anything in your own markup to use octothorpes as long as you can create a harmonizer that consistently identifies them. Harmonizer doc is TKTK as of 7.1.25 but you can see the default harmonizer in action here if you’d like to infer how they work.

Relevant RDF concepts

Chunks of data are referred to as “statements” in OP because the database is an RDF triplestore. It is hosted on an Oxigraph server which uses SPARQL queries to retrieve statements as JSON. The OP API provides a GET endpoint that builds dynamic SPARQL queries using routes and url params and returns them as structured JSON objects.

Every valid query is represented as a JSON object called a MultiPass, and the codebase has public functions that can accept Multipasses directly.

TKTK Multipass documentation

Subjects and Objects

Every triple has a subject and an object

The base statement in OP is that a subject octothorpes an object

“Hashtags” are when a subject octothorpes an object that is identified by a url for a Term on an OP server such as https://octothorp.es/~/demo

Since there are so many synonyms for “hashtag”, the technical label for them in OP is Term as they are completely separate RDF types from URLs that represent webpages. In the database it looks like this:

<https://octothorp.es/~/demo> rdf:Type <octo:Term>

whereas a subject that points to any other type of public URL is rdf:Type <octo:Page>

Octothorpe relationships to Terms are treated as hashtags and referred to simply as “octothorpes”

Octothorpe relationships to Pages are Links, which have subtypes such as Bookmarks

In the API, you filter for subjects and objects using the s and o url parameters (or not-s or not-o to exclude). Most endpoints require at one or more value, with some special cases requiring only one value.

URL Structure

https://octothorp.es/get/[what]/[by]/[[as]]

[what] - What type of data to retrieve
[by] - How to filter or match the data
[[as]] - Optional output format (defaults to JSON)

[what] has two formats for what it returns – blobjects or lists.

Everything = Blobjects

get/everything will return blobjects – hybrid, structured json objects that contain URI-level data for the pages returned as well as all of their octothorpes with their attached information, such as type (link, backlink etc)

Example: https://octothorp.es/get/everything/posted?s=demo

Specific things = lists

get/thorpes or get/pages will return a flat list of matching pages and data that only applies directly to their URL.

Example: https://octothorp.es/get/pages/posted?s=demo

Available Endpoints

For a continuously updated overview of the state of all the endpoints on the API, see this Google Sheet.

Data Types (`[what]`)

Type	Description	Returns
`everything`	Complete blobjects with all metadata	Blobjects (pages with octothorpes, links, etc.)
`pages`	List of pages/URLs	Pages List
`thorpes`	List of octothorpes/terms	Terms List
`domains`	List of domains	Pages List, domains only

Query Methods (`[by]`)

Method	Description	Subject Type	Object Type
`thorped`	Pages tagged with specific Terms	Originating Page	Term
`linked`	Pages that link to other pages	Originating Page	Linked Page
`backlinked`	Same as linked, scoped to validated backlinks	Originating Page	Page
`bookmarked`	Same as linked, scoped to subtype Bookmark	Originating Page	Linked Page
`posted`	Any Pages indexed (no object filter)	Originating Page	N/A
`in-webring`	Pages within a specific webring	Parent Webring (req)	Term (opt)

Output Formats (`[[as]]`)

Format	Description	Status
`json`	JSON response (default)	Available
`rss`	RSS 2.0 feed	Untested
`debug`	Debug information with query details	Available

Query Parameters

Core Parameters

Parameter	Type	Description	Default
`s`	String/Array	Subject URLs or terms to match	None (required)
`o`	String/Array	Object URLs or terms to match	None (required)
`match`	String	Matching strategy	`exact`

Filtering Parameters

Parameter	Type	Description	Default
`limit`	Integer	Maximum number of results	100
`offset`	Integer	Number of results to skip	0
`when`	String	Date/time filter	None

Matching Strategies (`match` parameter)

Strategy	Description	Match Strategy
`exact`	Exact URL/term matching	Uses the exact param as a SPARQL URI
`fuzzy-s`	Fuzzy matching only on subject	Uses param as a string to match records using a CONTAINS query only for subject (inclusion or exclusion), exact for object
`fuzzy-o`	Fuzzy matching only on object	Exact matches of common variations of terms as URIs (ie: hashtag, HashTag, #hashtag, hash tag), Exact for Subject
`fuzzy`	Fuzzy matching on both subject and object	Fuzzy match for both s and o
`very-fuzzy-o`	Very loose object matching	Matches common variations of terms using a CONTAINS query (can be slow)

Inferred matching

The API will infer some match formats based on the structure of the submitted values for s and o:

Well-formed URLs starting with https:// → exact matching
Strings → fuzzy matching
Match param value → Setting the match parameter overrides any applicable default

WARNING: At the moment, very-fuzzy matching will run very slowly when paired with a date filter.

Date Filtering (`when` parameter)

Format	Description	Example
`recent`	Last 2 weeks	`when=recent`
`after-DATE`	After specific date	`when=after-2024-01-01`
`before-DATE`	Before specific date	`when=before-2024-12-31`
`between-DATE-and-DATE`	Date range	`when=between-2024-01-01-and-2024-12-31`

Date formats accepted:

Unix timestamp: 1704067200
ISO date: 2024-01-01

Examples

Get all pages tagged with “demo”

https://octothorp.es/get/everything/thorped?o=demo

Get pages that link to any page on a certain domain

Note: using domain as a string rather than a well-defined url will run as a fuzzy match, returning results for all urls on that domain rather than that exact url

https://octothorp.es/get/pages/linked?o=demo.ideastore.dev

Get recent bookmarks with fuzzy matching

TKTK no good bookmark data right now. waiting for support of Terms on Bookmarks

This returns the MultiPass created from the request params, the literal SPARQL query generated from the MultiPass, and the actual results of the query in the following object:

{
  "results": {
    "multiPass": {},
    "query": {},
    "actualResults": {}
  }
}

Response Formats

Blobjects / Everything

{
  "results": [
    {
      "@id": "https://demo.ideastore.dev/multi-platform",
      "title": "Octothorpe Anywhere - Octothorpes Demo",
      "description": "Implementing Octothorpes is really easy. Here are some sites on other platforms that already have it.",
      "image": null,
      "date": 1740179856134,
      "octothorpes": [
        "demo",
        "wordpress"
      ]
    },
    {
      "@id": "https://demo.ideastore.dev/assertions",
      "title": "Assertions - Octothorpes Demo",
      "description": "Statements about urls other than the page you're on are called Assertions",
      "image": null,
      "date": 1740180203218,
      "octothorpes": [
        "etymology",
        "oof",
        "assertions",
        "octothorpe"
      ]
    },
    {
      "@id": "https://demo.ideastore.dev/page%20with%20spaces",
      "title": "Page with Spaces in URL - Octothorpes Demo",
      "description": "This page uses an alternate server, which can easily be set with a page variable.",
      "image": null,
      "date": 1740190020861,
      "octothorpes": [
        "testing",
        "demo"
      ]
    },
    ...
    ]
}

Pages List

{
  "results": [
    {
      "uri": "https://demo.ideastore.dev/multi-platform",
      "title": "Octothorpe Anywhere - Octothorpes Demo",
      "description": "Lorem!.",
      "date": 1740179856134,
      "image": null
    },
    {
      "uri": "https://demo.ideastore.dev/assertions",
      "title": "Assertions - Octothorpes Demo",
      "description": "Lorem ipsum",
      "date": 1740180203218,
      "image": null
    },
    {
      "uri": "https://demo.ideastore.dev/page%20with%20spaces",
      "title": "Page with Spaces in URL - Octothorpes Demo",
      "description": "Morem ipsum",
      "date": 1740190020861,
      "image": null
    }
  ]
}

Terms List

{
  "results": [
    {
      "term": "demo",
      "date": 1739540092164
    },
    {
      "term": "backlinks",
      "date": 1739540092166
    },
    {
      "term": "demo",
      "date": 1739547523260
    },
    {
      "term": "xoxofest",
      "date": 1739547523663
    },
    ...
    ]
  }

Debug Response

{
  "multiPass": {
    "meta": {
      "title": "Get blobjects matched by all (posted) as blobjects",
      "description": "MultiPass auto generated from a request to the https://octothorp.es//get API",
      "server": "https://octothorp.es/",
      "author": "Octothorpes Protocol",
      "image": "https://octothorp.es/badge.png",
      "version": "1",
      "resultMode": "blobjects"
    },
    "subjects": {
      "mode": "fuzzy",
      "include": [
        "demo"
      ],
      "exclude": []
    },
    "objects": {
      "type": "all",
      "mode": "exact",
      "include": [],
      "exclude": []
    },
    "filters": {
      "subtype": "",
      "limitResults": "100",
      "offsetResults": "0",
      "dateRange": {}
    }
  },
  // line breaks added to query here for readability
  "query": "SELECT DISTINCT
  ?s ?o ?title ?description ?image ?date ?pageType ?ot ?od ?oimg
      WHERE {    VALUES ?subList { \"demo\" }
        FILTER(CONTAINS(STR(?s), ?subList))
        ?s octo:indexed ?date .
        ?s rdf:type ?pageType .
        ?s octo:octothorpes ?o .
        OPTIONAL { ?s octo:title ?title . }
        OPTIONAL { ?s octo:image ?image . }
        OPTIONAL { ?s octo:description ?description . }
        OPTIONAL { ?o octo:title ?ot . }
        OPTIONAL { ?o octo:description ?od . }
        OPTIONAL { ?o octo:image ?oimg . }  }
        ORDER BY ?date",
  "actualResults": [
    {
      "uri": "https://demo.ideastore.dev/multi-platform",
      "title": "Octothorpe Anywhere - Octothorpes Demo",
      "description": "lorem ipsum.",
      "date": 1740179856134,
      "image": null
    },
    {
      "uri": "https://demo.ideastore.dev/assertions",
      "title": "Assertions - Octothorpes Demo",
      "description": "lorem ipsum",
      "date": 1740180203218,
      "image": null
    },
        ...
    ]
  }

Special Cases

Webrings: Always use exact matching for webring URLs
Terms: Default to exact matching for octothorpe terms

Error Handling

The API returns (some) appropriate HTTP status codes:

200 - Success
500 - Server error

TKTK:

400 - Invalid parameters
404 - No results found

Note: descriptive errors currently only report in dev console

Rate Limiting

Currently, the API has generous rate limits. Please be respectful and avoid making excessive requests.

CORS

The API supports CORS and can be used from web applications. All endpoints include the Access-Control-Allow-Origin: * header.

Versioning

This documentation covers API version 0.5 Breaking changes may occur until 1.0 but we’re trying to avoid that.

Support

For questions or issues with the API, please open an issue on Github or contact Nik or Nim directly.

# Octothorpes Demo

#